Computer Engineering Commons^™

Deep Learning And Localized Features Fusion For Medical Image Classification, Haidar A. Almubarak

Doctoral Dissertations

"Local image features play an important role in many classification tasks as translation and rotation do not severely deteriorate the classification process. They have been commonly used for medical image analysis. In medical applications, it is important to get accurate diagnosis/aid results in the fastest time possible.

This dissertation tries to tackle these problems, first by developing a localized feature-based classification system for medical images and using these features and to give a classification for the entire image, and second, by improving the computational complexity of feature analysis to make it viable as a diagnostic aid system in practical clinical …